ADVANCED MATERIALS FOR OPTICS AND ELECTRONICS

Adv[ Mater[ Opt[ Electron[ 09\ 112121 "1999#

Received 15 May 1999
Copyright 1999 John Wiley + Sons\ Ltd[ Accepted 10 June 1999
Substituent Effects on the Volatility of Metal
bb-diketonates
Bradley D. Fahlman
1
and Andrew R. Barron
1,2
*
1
Department of Chemistry and Center for Nanoscale Science and Technology, Rice University, Houston, Texas
77005, USA
2
Department of Mechanical Engineering and Materials Science, Rice University, Houston, Texas 77005, USA
Volatile trends are established for a series of M(bb-diketonate)
n
complexes, where M=Cu (n=2); Al, Sc,
Cr, Fe, Co, Ga (n=3); Zr (n=4) andbb-diketonate=acetylacetonate (acac), triuoroacetylacetonate (tfac),
hexauoroacetylacetonate (hfac) and 2,2,6,6-tetramethyl-3,5-heptanedionate (tmhd). Sublimation
enthalpies (DH
sub
) were calculated from thermogravimetric analysis (TGA) data, which show that the
dependence of DH
sub
on the number and type of intermolecular interactions appears to be more
substantial than molecular mass effects. Irrespective of the metal, the DH
sub
values for the M(tmhd)
3
derivative have a DH
sub
value of between 93 and 99% of that of the M(acac)
3
derivative. In contrast, the
M(tfac)
3
and M(hfac)
3
derivatives have DH
sub
values of ca. 82% and 54% of the values for the cor-
respondingM(acac)
3
. Similar trends are observedfor M(bb-diketonate)
2
andM(bb-diketonate)
4
derivatives.
The relationship between the DH
sub
and (T
sub
) of the substituted bb-diketonate derivatives as compared
withthe values for parent M(acac)
n
may be used topredict either quantity for a range of M(bb-diketonate)
n
complexes where the values for M(acac)
n
are known. Usingthe TGAsublimationdata, vapour pressures
have been calculated for each of the M(bb-diketonate)
n
complexes. Copyright 2000 John Wiley &
Sons, Ltd.
KEYWORDS b-diketonate; CVD precursor; enthalpy of sublimation; thermogravimetric analysis; vapour pres-
sure
INTRODUCTION
Metal b!diketonate chelate compounds are invalu!
able precursors for the chemical vapour deposition
"CVD# of metal and non!metal thin _lms[
0\1
In
general\ these compounds are favoured since they
are relatively volatile\ non!toxic and in the case of
the parent acetylacetonate complexes "I#\ inex!
pensive[
2
Furthermore\ replacing the substituent
methyl groups of the parent derivative with other
moieties has been demonstrated to in~uence the
Correspondence to] A[ R[ Barron\ Department of Chemistry\
Rice University\ 5099 Main Street\ Houston\ TX 66994\ USA[
e!mail] arbrice[edu
Contract grant sponsor] National Science Foundation[
Contract grant sponsor] O.ce of Naval Research[
volatility of these complexes[ A common method
used to enhance precursor volatility and cor!
responding e.cacy for CVD applications is to
incorporate partially "II# or fully "III# ~uorinated
ligands[
3
The observed enhancement in volatility may be
rationalized either by an increased amount of inter!
molecular repulsion due to the additional lone pairs
B[ D[ FAHLMAN AND A[ R[ BARRON
Copyright 1999 John Wiley + Sons\ Ltd[ Adv[ Mater[ Opt[ Electron[ 09\ 112121 "1999#
113
or that the reduced polarizability of ~uorine "rela!
tive to hydrogen# causes ~uorinated ligands to have
less intermolecular attractive interactions[
4
Caulton
and co!workers
2
have determined that\ for a series
of ~uorinated zirconium ~uoroalkoxides\ the
degree of volatility enhancement is a function of
the degree of ligand ~uorination[ For example\
Na"OR
f
#
3
species showed a signi_cant increase in
volatility as the ~uorine content was increased[
Although the monomeric Zr"OR
f
#
3
complexes dis!
played an analogous behaviour\ it was to a much
smaller extent[ Similarly\ in determining the
thermodynamic properties for various complexes\
Ni"salen#Ln"hfa#
2
"LnY\ Gd#\ Gleizes et al[
5
showed that solids possessing similar molecular and
solid!state structures may have di}erent volatilities
depending on the nature of the ligand and the
dimerization propensity of the complex[
A prediction of the volatility of a metalorganic
compound as a function of its ligands and molecular
structure would be desirable in order to determine
the suitability of such a compound as a CVDprecur!
sor[ In a study dealing with a range of inorganic and
organic compounds
6
it was determined that a rough
proportionality exists between a compound|s melt!
ing point and its sublimation enthalpy[ However\
the largest deviations from the reported correlations
were found for the inorganic compounds that were
investigated[ Hence\ this reported correlation
appears to be an oversimpli_cation that may be best
limited to simple organic solids\ not inorganic or
organometallic complexes[ In this regard\ we have
undertaken a determination of sublimation enthalp!
ies for a range of metal b!diketonate complexes in an
attempt to correlate volatility with solid!state and:or
molecular structures[
Enthalpies of sublimation for metal b!diketonate
chelates have been previously determined through a
variety of methods\ most commonly from vapour
pressure measurements using complex experimental
systems such as Knudsen e}usion\
7
temperature
drop microcalorimetry
8
and\ more recently\ di}er!
ential scanning calorimetry "DSC#[
09\00
However\ the
measured values are highly dependent on the exper!
imental procedure utilized[ For example\ the
reported sublimation enthalpy of Al"acac#
2
"I\ where
MAl\ n2# varies from 36[2 to 015 kJ mol
0
[
01
While other thermal techniques\ such as ther!
mogravimetric analysis "TGA#\ have been utilized in
molecular decomposition studies\
02\03
the few reports
using such an approach to determine sublimation
enthalpies have provided relatively accurate and pre!
cise values[
04\05
It is this method that we have
employed to examine the thermochemical properties
for selected metal b!diketonates[
RESULTS AND DISCUSSION
A series of metal b!diketonate complexes were
obtained from commercial sources or prepared
according to literature methods[ Each compounds
was puri_ed by recrystallization or sublimation and
used as a polycrystalline solid for thermogravi!
metric:di}erential thermal analyses "TG:DTA#\ see
Experimental[ The non!~uorinated "I\ acetylace!
tonate\ acac#\ tris!~uorinated "II\ tri~uoroacetyl!
acetonate\ tfac# and hexakis!~uorinated "III\
hexa~uoroacetylacetonate\ hfac# series of ligands
were chosen to provide a measure of the e}ect on
volatility with ~uorine substitution[ In addition\ the
tert!butyl substituted ligand "IV\ 1\1\5\5!tetra!
methyl!2\4!heptanedionate\ tmhd# was used to
determine the e}ect of added steric bulk on vola!
tility[
Prior to a detailed analysis of volatility\ it was
important to _rst assess the volatility of each b!
diketonate complex over a wide temperature range
to ensure that each compound could be sublimed
intact without decomposition\ i[e[\ equation "0#[
M"b!diketonate#
n

"solid#
:M"b!diketonate#
n

"vapour#
"0#
A simultaneous TG:DTA instrument was used to
monitor the sample mass losses and energetic events
as a function of temperature[ The uniform mass
loss of M"b!diketonate#
n
is accompanied by a broad
endothermic event characteristic of sublimation[
06
In some compounds this thermal event is quite
broad or is shifted near the end of the sublimation
process\ while in other cases melting occurs as evi!
VOLATILITY OF METAL b!DIKETONATES
Copyright 1999 John Wiley + Sons\ Ltd[ Adv[ Mater[ Opt[ Electron[ 09\ 112121 "1999#
114
Table 0[ Thermodynamic results on metal b!diketonate compounds
Measured T
sub
Vapour
M
w
mp range DH
sub
DS
sub
calculated T
19
pressure at
Compound "g mol
0
# ">C# ">C# "kJ mol
0
#
$
"J K
0
mol
0
# ">C# ">C# at 049 >C "Torr#
Al"acac#
2
213 081 039069 82 "008#
%
119 049 060 2[150
Al"tfac#
2
375 013 099029 63 "090#
%
081 000 021 8[604
Al"hfac#
2
537 61 5989 41 "61#
%
041 69 80 18[019
Al"tmhd#
2
466 145 039069 77 102 039 050 2[267
Sc"acac#
2
231 194 039069 84 113 040 061 2[107
Sc"tfac#
2
493 071 099029 67 192 001 021 8[481
Sc"hfac#
2
555 89 5989 44 048 62 83 17[789
Sc"tmhd#
2
484 040 039069 89 107 031 051 2[220
Cr"acac#
2
238 103 039069 80 "012#
10
105 037 056 2[217
Cr"tfac#
2
400 074 099029 60 "006#
10
075 098 017 8[809
Cr"hfac#
2
562 73 5989 35 "001#
10
023 58 78 18[400
Cr"tmhd#
2
591 120 039069 74 196 026 045 2[321
Fe"acac#
2
242 073 039069 001 "002#
19
148 050 071 1[670
Fe"tfac#
2
404 000 099029 85 132 010 032 7[239
Fe"hfac#
2
566 86 5989 59 058 70 090 14[910
Fe"tmhd#
2
595 053 039069 000 152 049 066 1[809
Co"acac#
1
146 069 059089 038 217 070 199 9[714
Co"tmhd#
1
315 143 059089 032 211 060 081 9[787
Co"acac#
2
245 100 059089 027 "007#
10
200 069 089 0[948
Co"tfac#
2
407 044 099029 008 "003#
10
184 020 042 2[208
Co"hfac#
2
579 83 5989 62 199 89 001 8[021
Co"tmhd#
2
598 032 059089 021 294 050 068 0[100
Cu"acac#
1
151 173 039069 019 "005#
28
165 052 088 0[702
Cu"tfac#
1
258 084 099029 001 "001#
10
171 013 038 1[881
Cu"hfac#
1
& 367 74 5989 86 "097#
10
176 70 78 3[934
Cu"tmhd#
1
329 087 039069 003 162 044 078 0[883
Ga"acac#
2
256 082 039069 89 103 036 057 2[228
Ga"tfac#
2
418 039 099029 64 088 097 016 09[920
Ga"hfac#
2
580 72 5989 42 046 57 77 29[004
Ga"tmhd#
2
508 119 039069 76 102 027 046 2[348
Zr"acac#
3
377 061 039069 015 178 051 074 0[608
Zr"tfac#
3
693 015 099029 83 132 005 039 8[609
Zr"hfac#
3
808 31 5989 48 055 72 094 13[789
Zr"tmhd#
3
713 208 039069 019 167 048 079 0[761
Melting points are from Strem Chemicals\ Inc[
$
Literature values given in parentheses[
%
Corresponds to the enthalpy of vaporization[
01
&
Cu"hfac#
1
[xH
1
O "DH
sub
091[9 "kJ mol
0
#\ mp 87 ">C#[
denced by a sharp endotherm[ Therefore\ the pos!
ition of this thermal event was not a good indicator
of relative volatilities[ In order to allow a simple
and direct comparison of the relative volatility of
the various compounds\ a sublimation temperature
was de_ned as the point when 19) mass loss had
occurred "i[e[\ T
19
# and T
19
values were determined
for each compound under ~owing inert gas "argon#]
these values are listed in Table 0[
Determination of Sublimation Enthalpies
(DH
sub
) of Metal b-diketonates
The enthalpy of sublimation is a quantitative mea!
sure of the volatility of a particular solid[ This infor!
mation is useful when considering the feasibility of
a particular precursor for CVD applications[ An
ideal sublimation process involves no compound
decomposition and only results in a solidgas phase
B[ D[ FAHLMAN AND A[ R[ BARRON
Copyright 1999 John Wiley + Sons\ Ltd[ Adv[ Mater[ Opt[ Electron[ 09\ 112121 "1999#
115
Fig[ 0[ Plot of TGA results for Cr"acac#
2
performed at di}erent
isothermal regions
change "i[e[\ equation "0##[ Since phase changes are
thermodynamic processes following zero!order kin!
etics\ the evaporation rate or rate of mass loss by
sublimation "m
sub
#\ at a constant temperature "T#\
is constant at a given temperature "equation "1##[
m
sub

Dmetalb!diketonate
Dt
"1#
Therefore\ the m
sub
values may be directly deter!
mined from the linear mass loss of the TGA data
in isothermal regions[ As an illustrative example\
Fig[ 0 displays the data for the mass loss of Cr"a!
cac#
2
at three isothermal regions under a constant
argon ~ow[ Each isothermal data set exhibits a lin!
ear relation "R
1
9[88#[ As expected for an endo!
thermal phase change\ the linear slope\ equal to
m
sub
\ increases with increasing temperature[
The ClausiusClapeyron relation between pres!
sure "p# and temperature "T#\
07
shown in equation
"2#\ forms the basis of analysing isothermal TGA
data[
d ln"p#
dT

DH
sub
RT
1
"2#
Since m
sub
data are obtained from TGA data\ it is
necessary to utilize the Langmuir equation "3#
which relates the vapour pressure of a solid with its
sublimation rate[
08
Fig[ 1[ Plot of log"m
sub
zT# versus 0:T and the determination of
the DH
sub
"001[5 kJ mol
0
# for Fe"acac#
2
"R
1
9[8878#
p
$
1pRT
M
w
%
9[4
m
sub
"3#
After integrating equation "2# in log form and sub!
stituting equation "3#\ one obtains the equality given
by equation 4[
04
log"m
sub
zT#
9[9411"DH
sub
#
T

$
9[9411"DH
sub
#
T
sub

0
1
log
0
0295
M
w 1%
"4#
Hence\ the linear slope of a log"m
sub
zT# versus 0:T
plot yields DH
sub
[ An example of a typical plot and
the corresponding DH
sub
value is shown in Fig[ 1[
In addition\ the y intercept of such a plot provides
a value for T
sub
\ the calculated sublimation tem!
perature at atmospheric pressure[
Table 0 lists the results using the TGA method
for a variety of metal b!diketonates[ The value
determined for Fe"acac#
2
shows a di}erence of less
than 3) relative to the literature value
19
and is
within experimental error of the value obtained by
the analogous method on a di}erent instrument
reported by Gillan et al[
04
However\ the values
determined for M"acac#
2
\ M"tfac#
2
and M"hfac#
2
"MAl\ Cr#\ and Co"acac#
2
di}er from previously
reported values[
01\10
This may be rationalized by
examining the methods utilized for each compound[
VOLATILITY OF METAL b!DIKETONATES
Copyright 1999 John Wiley + Sons\ Ltd[ Adv[ Mater[ Opt[ Electron[ 09\ 112121 "1999#
116
For Cr"acac#
2
and Co"acac#
2
\ isoteniscopic and sub!
limation bulb techniques were used\ whose accu!
racy have long been questioned[
19
For the
aluminium b!diketonate complexes\ a comparable
thermogravimetric approach was employed\
01
but\
their system used bu}er gas pressures in excess of
099 Torr[ Since the pressure of their system was
su.ciently greater than that used to obtain the
values listed in Table 0\ it is no surprise that a
su.cient discrepancy exists between the two tech!
niques[ However\ since both methods illustrate the
same trend in DH
sub
with increasing ligand ~uo!
rination\ the data o}set is most likely the direct
consequence of di}ering system pressures[ Since the
system used in our experiments provides the same
data that we have obtained using a dynamic
vacuum\ it is unlikely that a systematic error is
being introduced into the numbers listed in Table 0
through the utilization of a steady ~ow of Ar gas[
Furthermore\ for all metal complexes\ the order of
volatility follows the trend acc tfac hfac which
agrees with the order reported for Cu"II#
complexes[
11
The entropy of sublimation is readily calculated
from the DH
sub
and the calculated T
sub
data in Table
0\ equation "5#[ The range observed for all the metal
b!diketonates compounds DS
sub
029229 J K
0
mol
0
would be expected for a transformation giv!
ing translational and internal degrees of freedom[
Interestingly\ for any particular metal the lowest
DS
sub
are observed for the ~uorinated derivatives\
see Table 0[
DS
sub

DH
sub
T
sub
"5#
While the sublimation temperature is an important
parameter to determine the suitability of potential
precursor compounds for CVD\ it is often pref!
erable to express a compound|s volatility in terms of
its vapour pressure[ However\ while it is relatively
straightforward to determine the vapour pressure
of a liquid or gas\ measurements of solids is di.cult
"e[g[\ use of the isoteniscopic method#[
12
Sub!
stitution of equation "3# into equation "4# allows
for the calculation of the vapour pressure "p# of the
metal b!diketonates compounds as a function of
temperature "T#[ The calculated vapour pressures
for each of the metal b!diketonates compounds at
049 >C are given in Table 0[ We have found the
TGA approach to show reasonable agreement with
previous measurements[ For example\ the value cal!
culated for Fe"acac#
2
"e[g[\ 1[67 Torr at 002 >C# is
slightly higher than that measured directly by the
isoteniscopic method "9[42 Torr at 002 >C#\
however\ it should be noted that measurements
using the sublimation bulb method obtained values
much lower "709
2
Torr at 002 >C#[
13
Based on
these results\ we propose that this TGA method
o}ers a suitable alternative to conventional "direct#
measurements of vapour pressure[
Structural Correlations Between Calculated
DH
sub
Values
The measured DH
sub
values for M"acac#
2
and
M"tmhd#
2
complexes indicate that replacing the
methyl groups with tert!butyl moieties slightly
increases the volatility of the complex[ Although it
would be expected that the increased mass i[e[\
Al"acac#
2
213[20 g mol
0
versus Al"tmhd#
2

465[79 g mol
0
should decrease the overall vola!
tility of the complex\ the relatively low enthalpy
values indicate that the CH = = = HC interactions
between neighbouring tert!butyl groups is approxi!
mately equal in magnitude to methyl = = = methyl
interactions[ In all cases\ the tmhd derivative has a
DH value of between 82 and 88) of that of the
acac derivative[ This e}ect is irrespective of the
number of the oxidation state of the metal\ i[e[\ the
number of b!diketonate ligands[
As was expected based upon previous workers|
results\ the ~uorinated complexes are more volatile
than the non!~uorinated derivatives[ Furthermore\
the higher the ~uorination "i[e[\ hfac versus tfac
versus acac# the greater the increase in volatility[
This trend is expected based solely upon the di}er!
ence in intermolecular CH = = = HC versus CH
= = = FC versus CF = = = FC interactions[ Fig[ 2"a#
represents a comparison of the DH
sub
for M"tfac#
n
and M"hfac#
n
normalized to M"acac#
n
[ This shows
that replacement of one CH
2
"per ligand# with CF
2
results in a decrease in DH
sub
of ca[ 72) for M"b!
diketonate#
2
irrespective of the identity of the metal[
Replacing the second CH
2
results in another 55)
decrease in the DH
sub
[ It may be clearly seen from
Fig[ 2"a# that the amount of decrease is dependent
on the oxidation state of the metal\ i[e[\ more
ligands cause a larger change in volatility[ A similar
trend is observed for T
sub
"Fig[ 2"b##^ however\ there
appears to be no e}ect by changing the number of
ligands "oxidation state#[
B[ D[ FAHLMAN AND A[ R[ BARRON
Copyright 1999 John Wiley + Sons\ Ltd[ Adv[ Mater[ Opt[ Electron[ 09\ 112121 "1999#
117
Fig[ 2[ Plot of relative "a# DH
sub
and "b# T
sub
\ normalized to
M"acac#
n
\ versus the number of b!diketonate ligands "n#^ acac
"#\ tfac "# and hfac "#
We propose that the relationship between the
DH
sub
and T
sub
of the ~uorinated derivatives as com!
pared with the values for M"acac#
n
may be used to
predict either quantity for a range of M"tfac#
n
and
M"hfac#
n
complexes where the values for the parent
M"acac#
n
are known[ In fact\ with similar trends it
should be possible to determine these values for a
wide range of b!diketonate complexes[ An inter!
esting application of these relationships is that the
deviation from the expected trend may be used to
con_rmthe presence or absence of oligomers for the
sterically undemanding complexes\ i[e[\ Ba"acac#
1
is
known to exist as an oligomer in the solid and
vapour state and hence has a signi_cantly lower
volatility than M"tmhd#
1
[ Finally\ the DH
sub
and
T
sub
may be indirectly determined for compounds
that decompose during sublimation[
The tris!acetylacetonate complexes of Al"III#\
14
Cr"III#\
15
Co"III#\
16
and Ga"III#
17
forma monoclinic
isomorphic set of solid state structures\ crystallizing
in the P1
0
:c space group[ By comparison\ Fe
"acac#
2
18
and Sc"acac#
2
29
each possess orthorhombic
structures\ in the Pbca space group[ Since the struc!
tures of the Al\ Cr\ Co and Ga acetylacetonate
complexes are isomorphous\ one would expect the
volatility to be related to the molecular weight of
the complex\ since no factors such as di}ering metal
oxidation states or ~uorinated ligands exist[ Fig[ 3
shows plots of sublimation enthalpies for "a#
M"acac#
2
\ "b# M"tmhd#
2
\ "c# M"tfac#
2
and "d#
M"hfac#
2
as a function of molecular weight[ Two
di}erent trends between volatility and molecular
weight are observed^ d
n
transition metal complexes
exhibit a linear correlation whereas no correlation
between molecular weight and volatility is evident
for d
9
metal complexes\ see Fig[ 3"a#[ While Gillan
et al[
04
discovered a link between volatility and
molecular mass\ they also found that the degree of
branching of alkyl groups was more in~uential than
molecular mass on the volatility of a series of gal!
lium chalcogenide cubanes\ presumably due to
stronger intermolecular interactions[ Since the simi!
lar volatilities between M"acac#
2
and M"tmhd#
2
complexes also indicate that molecular weight is
relatively unimportant\ the nature and degree of the
intermolecular interactions are likely to be pre!
dominant factors in the volatilities for these com!
plexes[
Table 1 lists the closest intermolecular inter!
actions and number of neighbouring molecules for
M"acac#
2
complexes[ DH
sub
values for the M"acac#
2
complexes indicate that Co"acac#
2
and Fe"acac#
2
should have the most intermolecular interactions
giving rise to the lowest volatility[ For Co"acac#
2
\
the closest interactions are the CH = = = O "C = = =
O237 _# which is lower than the sum of the Van
der Waals| radii "ca[ 2[4 _
20
#[ No such interaction
exists for the Fe"acac#
2
complex\ as the closest C
H = = = O distance is 2[845 _^ outside the sum of the
Van der Waals| radii[ However\ in Fe"acac#
2
\ C
H = = = HC "methyl [ [ [ methyl and methyl [ [ [ meth!
VOLATILITY OF METAL b!DIKETONATES
Copyright 1999 John Wiley + Sons\ Ltd[ Adv[ Mater[ Opt[ Electron[ 09\ 112121 "1999#
118
Fig[ 3[ Plot of DH
sub
versus molecular weight for "a# M"acac#
2
\ "b# M"tmhd#
2
\ "c# M"tfac#
2
and "d# M"hfac#
2
complexes
ine# are the predominant interactions and much
stronger in magnitude than the analogous contacts
in Co"acac#
2
[ Hence\ the lack of CH = = = O inter!
actions in the Fe"acac#
2
complex relative to the
cobalt analogue most likely contributes to the
observed greater volatility[ The intermolecular
interactions for Cr"acac#
2
\ Al"acac#
2
\ Sc"acac#
2
and
Ga"acac#
2
are all comparable "see Table 1# which
would correspond to the similarity in the measured
DH
sub
values for these complexes[ The major inter!
molecular contacts are the CH = = = HC inter!
actions since the CH = = = O distances are all greater
than the sum of the Van der Waals| radii[ Although
Sc"acac#
2
crystallizes in the same space group as
Fe"acac#
2
\ the CH
2
= = = HC distance is much shorter
for Fe"acac#
2
\ which must account for the sig!
ni_cantly lower volatility of Fe"acac#
2
[ Fig[ 4 illus!
trates the relationship between the shortest
intermolecular interaction "below the sum of the
appropriate van der Waal radii# and the DH
sub
[ A
general trend is observed of decreasing inter!
molecular contacts with decreasing volatility for
B[ D[ FAHLMAN AND A[ R[ BARRON
Copyright 1999 John Wiley + Sons\ Ltd[ Adv[ Mater[ Opt[ Electron[ 09\ 112121 "1999#
129
Table 1[ Comparison of the shortest intermolecular contacts for
M"acac#
2
complexes
Number of CH= = = O CH= = = HC$ CH= = = C%
nearest "A# "A# "A#
Compound neighbours
Al"acac#
2
3 2[467 2[749 2[654
Sc"acac#
2
5 2[719 2[569 2[769
Cr"acac#
2
3 2[405 2[801 2[783
Fe"acac#
2
5 2[845 2[558 2[585
Co"acac#
2
3 2[347 2[748 2[722
Ga"acac#
2
3 2[476 2[789 2[663
The sum of the Van der Waals| radii for CH = = = O and
CH = = = HC are ca[ 2[4 _ and 3[9 _\ respectively[
$
Methyl [ [ [ methyl interaction[
%
Methyl [ [ [ methine interaction[
Fig[ 4[ Plot of the DH
sub
values versus the closest intermolecular
contact distances "_# for M"acac#
2
complexes illustrating the
general trend of increasing volatility with increasing inter!
molecular contacts[ The sum of the Van der Waals| radii for C
H = = = O and CH = = = HC are shown
M"acac#
2
complexes[ For Co"acac#
2
\ the short CH
= = = O distance appears to be most responsible for
the signi_cantly lower volatility exhibited by this
complex[ For all other compounds\ the CH = = =
HCinteractions are predominant^ their variance is
roughly in accord with the measured DH
sub
values[
The large inductive e}ect created by CF
2
groups
should give rise to a signi_cantly di}erent charge
distribution in the chelate ring and a cor!
respondingly di}erent co!ordination sphere in ~u!
orinated derivatives relative to acac complexes[
21
However\ the co!ordination environments in
Al"tfac#
2
\
22
Co"tfac#
2
\
22
Fe"hfac#
2
\
23
Cu"hfac#
1
24\25
and Cu"tmhd#
1
26
are essentially isostructural to
their acac counterparts[ Hence\ as expected\ the
same overall volatility trends previously described
for M"acac#
2
complexes are observed for the tfac
and hfac analogues "Fig[ 3#[
CONCLUSIONS
The sublimation enthalpies "DH
sub
# and sub!
limation temperatures "T
sub
# for a series of metal b!
diketonate complexes\ M"b!diketonate#
n
\ were
determined using thermogravimetric analysis[
Replacing methyl groups in M"acac#
n
with tert!
butyl groups in M"tmhd#
n
had little e}ect on the
volatility "although the e}ect is to generally increase
the volatility#[ In contrast\ replacement of methyl
with CF
2
units signi_cantly increased the volatility
in the order M"acac#
n
M"tfac#
n
M"hfac#
n
[ The
relationship between this homologous series may
be used to predict the DH
sub
and:or T
sub
for
unknown members of the series[
For the d
n
transition metal\ tris!b!diketonate
complexes investigated\ the increase in volatility of
a series of homologous compounds was observed
to increase linearly with the molecular weight of the
complex[ By contrast\ the volatility of analogous d
9
metal chelates was independent of molecular weight
and the degree and nature of intermolecular inter!
actions[ Investigation of the closest intermolecular
interactions shows that this appears to be the con!
trolling factor in determining the sublimation
enthalpies "DH
sub
#[
EXPERIMENTAL SECTION
Compounds Sc"hfac#
2
\ Cr"hfac#
2
\ Fe"hfac#
2
\
Co"hfac#
2
\ Co"tmhd#
2
\ Cu"hfac#
1
[xH
1
O Zr"tfac#
3
and Zr"hfac#
3
were prepared by the literature
method[
27
All other metal b!diketonate complexes
were purchased "Strem or Aldrich# and were pur!
i_ed by recrystallization in acetone or by vacuum
sublimation prior to use[ Compound Cu"hfac#
1
was
synthesized by sublimation of Cu"hfac#
1
[xH
1
O
under dynamic vacuum at 5969 >C[ Upon
exposure to air\ the blue anhydrous crystals rapidly
turned green due to hydrate formation[ Although
all compounds are available commercially\ there
VOLATILITY OF METAL b!DIKETONATES
Copyright 1999 John Wiley + Sons\ Ltd[ Adv[ Mater[ Opt[ Electron[ 09\ 112121 "1999#
120
Table 2[ Multinuclear NMR characterization of synthesized
metal b!diketonates
$
Compound
0
H
02
C
08
F
Al"tfac#
2
4[52 "CH
2
# 080[1 "COCH
2
# 64[78
0[36 "CH# 062[1 "COCF
2
#
006[1 "CF
2
#
86[5 "CH#
17[0 "CH
2
#
Al"hfac#
2
9[83 "CH# 070[1 "CO# 64[89
005[5 "CF
2
#
83[7 "CH#
Cr"hfac#
2
9[81 "CH# 079[8 "CO# 64[61
004[8 "CF
2
#
83[9 "CH#
Fe"hfac#
2
9[78 "CH# 079[0 "CO# 64[07
004[5 "CF
2
#
82[1 "CH#
Co"tfac#
2
4[40 "CH
2
# 077[8 "COCH
2
# 64[53
0[17 "CH# 060[9 "COCF
2
#
005[9 "CF
2
#
15[8 "CH
2
#
83[4 "CH#
Cu"tfac#
1
4[11 "CH
2
# 089[0 "COCH
2
# 65[94
0[21 "CH# 057[4 "COCF
2
#
005[6 "CF
2
#
85[1 "CH#
16[4 "CH
2
#
Cu"hfac#
1
9[80 "CH# 072[0 "CO# 65[24
006[0 "CF
2
#
84[4 "CH#
Zrtfac#
3
4[62 "CH
2
# 089[7 "COCH
2
# 64[27
0[44 "CH# 061[7 "COCF
2
#
005[8 "CF
2
#
84[1 "CH#
16[3 "CH
2
#
Chemical shifts "d# in ppm[
$
Assignments given in parentheses[
are incomplete reports of their NMR charac!
terization in the literature[ Hence\
08
F\
02
C and
0
H
NMRcharacterizations were performed on the syn!
thesized complexes "Table 2# using a Bruker 149
MHz and 399 MHz multinuclear spectrometer[
Thermogravimetric analyses experiments were
performed on a TA Instruments Simultaneous
Di}erential Techniques "SDT 1859# TGA!DTA
system using the methods previously described[
04
The argon ~ow rate was set to 89[9 ml min
0
and
was carefully monitored to ensure a steady ~owrate
during runs and an identical ~ow rate from one set
of data to the next[ Samples of iron acetylacetonate
was used as a calibration standard through DH
sub
determinations before each day of use[ If the mea!
sured value of the sublimation enthalpy for Fe"a!
cac#
2
was found to di}er from the literature value
by more than 4)\ the sample was reanalysed and
the ~ow rates were optimized until an appropriate
value was obtained[ Only after such a calibration
was optimized would other metal complexes be ana!
lysed in this manner[ As the surface area of a given
crystallite decreases during the experiment\ the sub!
limation rate will also decrease\ therefore\ to ensure
that the solid displayed a near constant surface
area\ data were taken over small temperature ranges
"ca[ 2939 >C# and the overall mass loss was kept
relatively low[ This latter condition was imposed
on the experiments since we have previously found
that 14) overall mass loss corresponded to a
change in surface area of less than 04)[
04
The mass
loss versus time curves were monitored for the com!
pounds being investigated to ensure linearity was
maintained[ Any nth order deviations for these
trends would correspond to decomposition pro!
cesses for the solids being investigated[ Also\
through the utilization of a simultaneous TG:DTA
system\ it would have been possible to observe such
exothermic events for a particular solid[ No such
features were observed for the compounds over the
temperature ranges investigated[ All the com!
pounds investigated exhibited full sublimation:
vaporization\ leaving less than 2) residue in most
cases[ Since the enthalpies of sublimation were to
be compared\ the isothermal ranges investigated
were those below the melting points of the com!
pounds studied "see Table 0#[ For the M"hfac#
n
com!
plexes this required that temperature ranges close
to ambient were used for the analyses[
ACKNOWLEDGEMENTS
Financial support for this work was provided by
the National Science Foundation and the O.ce of
Naval Research[
REFERENCES
0[ Rees WS Jr[ CVD of Nonmetals[ VCH] New York\
0885[
1[ Kodas T\ Hampden!Smith M[ The Chemistry of
Metal CVD[ VCH] New York\ 0885[
2[ Samuels JA\ Folting K\ Hu}man JC\ Caulton KG[
Chem[ Mater[ 0884^ 6] 818[
B[ D[ FAHLMAN AND A[ R[ BARRON
Copyright 1999 John Wiley + Sons\ Ltd[ Adv[ Mater[ Opt[ Electron[ 09\ 112121 "1999#
121
3[ See\ for example\ "a# Je}eries PM\ Wilson SR\ Gir!
olami GS[ Inor`[ Chem[ 0881^ 20] 3492^ "b# Bradley
DC\ Chudzynska H\ Hammond ME\ Hursthouse
MB\ Motevalli M\ Ruowen W[ Polyhedron 0881^ 00]
264^ "c# Purdy AP\ Berry AD\ Holm RT\ Fatemi
M\ Gaskill DK\ Inor`[ Chem[ 0878^ 17] 1688^ "d#
Thompson SC\ Cole!Hamilton DJ\ Gilliland DD\
Hitchman ML\ Barnes JC[ Adv[ Mater[ Optics Elec!
tron 0881^ 0] 70^ "e# Belcher R\ Cranley CR\ Majer
JR\ Stephen WI\ Uden PC[ Anal[ Chim[ Acta 0861^
59] 098[
4[ Reed TM III[ In Fluorine Chemistry\ Vol[ 4\ Simons
JH "ed#[ Academic Press] New York\ 0853^ 022[
5[ Gleizes A\ Julve M\ Kuzimina N\ Alikhanyan A\
Lloret F\ Malkerova I\ Sanz JL\ Senoca F[ Eur[ J[
Inor`[ Chem[ 0887^ 7] 0058[
6[ Westwell MS\ Searle MS\ Wales DJ\ Williams DH[
J[ Am[ Chem[ Soc[ 0884^ 006"07#] 4902[
7[ Hill JO\ Murray JP[ Rev[ Inor`[ Chem[ 0882^ 02] 014[
8[ Beech G\ Lintonbon RM[ Thermochim[ Acta 0860^
2] 86[
09[ Murray JP\ Cavell KJ\ Hill JO[ Thermochim[ Acta
0879^ 25] 86[
00[ Torres!Gomez LA\ Barreiro!Rodriquez G\ Galarza!
Mondragon A[ Thermochim[ Acta 0877^ 013] 118[
01[ Pauleau Y\ Dulac O[ Chem[ Mater[ 0880^ 2"1#] 170[
02[ Bukovec N\ Bukovec P\ Siftar J[ Thermochim[ Acta
0879^ 24] 74[
03[ Sli_rski J\ Huchet G\ Marty A\ Teyssandier F[ Chem[
Mater[ 0884^ 6] 511[
04[ Gillan EG\ Bott SG\ Barron AR[ Chem[ Mater[ 0886^
8"2#] 685[
05[ Ashcroft SJ[ Thermochim[ Acta 0860^ 1] 401[
06[ Turi EA\ Khanna YP\ Taylor TJ[ In A Guide to
Materials Characterization and Analysis\ Sibilia JP
"ed#[ VCH] New York\ 0877^ chapter 8[
07[ Atkins PW[ Physical Chemistry\ 4th edn[ W[H[ Free!
man] New York\ 0883^ 087[
08[ Langmuir I[ Phys[ Rev[ 0802^ 1] 218[
19[ Sabbah R\ Tabet D\ Belaadi S[ Thermochim[ Acta
0883^ 136] 082[
10[ Ribeiro da Silva MAV\ Ferrao MLCCH[ Pure and
Appl[ Chem[ 0877^ 59"7#] 0114[
11[ Baidina IA\ Sabnikov PA\ Igumenov IK\ Borisov
SV[ Koord[ Khim[ 0873^ 09] 837[
12[ Arm H\ Daeniker H\ Schaller R[ Helv[ Chim[ Acta
0855^ 37] 0661[
13[ Sachinidis J\ Hill JO[ Thermochimica Acta 0879^ 24]
48[
14[ "a# Hon PK\ P~uger CE[ J[ Coord[ Chem[ 0862^ 2^
56^ "b# Rahman A\ Ahmed SN\ Khair MA\ Zan!
grando E\ Randaccio L\ J[ Ban`ladesh Acad[ Sci
0889^ 03] 050[
15[ Morosin B[ Acta Crystallo`r[ 0854^ 08] 020[
16[ Kruger GJ\ Reynhardt EC[ Acta Cryst[ 0863^ B29]
711[
17[ Dymock K\ Palenik GJ[ Acta Cryst[ 0863^ B29] 0253[
18[ Roof RB[ Acta Cryst[ 0845^ 8] 670[
29[ Anderson TJ\ Newman MA\ Melson GA[ Inor`[
Chem[ 0862^ 01] 816[
20[ Handbook of Chemistry and Physics\ 59th edn[ CRC
Press] Boca Raton\ 0879[
21[ Watkins CL\ Harris ME[ J[ Inor`[ Nucl[ Chem[ 0867^
39] 0658[
22[ Vogelson CT\ Edwards CL\ Kobylivker AN\ Chacko
SB\ Moran CE\ Dalton K\ Stewart SM\ Werner BC\
Bott SG\ Barron AR[ J[ Chem[ Cryst[ 0887^ 17"00#]
704[
23[ P~uger CE\ Haradem PS[ Inor`[ Chim[ Acta 0872^
58] 030[
24[ Jain A\ Kodas TT\ Corbitt TS\ Hampden!Smith MJ[
Chem[ Mater[ 0885^ 7] 0008[
25[ Pinkas J\ Hu}man JC\ Baxter DV\ Chisholm MH\
Caulton KG[ Chem[ Mater[ 0884^ 6] 0478[
26[ Patnaik S\ Guru Row TN\ Raghunathan L\ Devi A\
Goswami J\ Shivashankar SA\ Chandrasekaran S\
Robinson WT[ Acta Cryst[ 0885^ C41] 780[
27[ Morris ML\ Moshier RW\ Sievers RE[ Inor`[ Chem[
0852^ 1] 300[
28[ Ribeiro da Silva MAV\ Ferrao MLCCH[ J[ Chem[
Thermodyn[ 0883^ 15] 204[